Facial expression recognition method based on deep convolutional neural network combined with improved LBP features

Fanzhi Kong
DOI: https://doi.org/10.1007/s00779-019-01238-9
2019-06-08
Personal and Ubiquitous Computing
Abstract:Aiming at the disadvantages of the traditional machine-based facial expression recognition method that eliminates the feature of manual selection, a feature extraction method based on deep convolutional neural network to learn expression features is proposed. Since the deep convolutional neural network can directly use the original image as the input image, the image abstract feature interpretation is obtained at the fully connected layer of the image, which avoids the inherent error of image preprocessing and artificial selection features. Then, we reconstruct the traditional local binary pattern (LBP) feature operator for facial expression image and fuse the abstract facial expression features learned by the deep convolution neural network with the modified LBP facial expression texture features in the full connection layer. A new facial expression feature can be obtained, and the classification accuracy can be improved. In general, for the recognition of facial expression images, the proposed method based on the fusion LBP expression features and convolutional neural network expression features is used to obtain the best performance of 91.28% in the comparative experiment. An efficient extension of the expression feature texture expression channel is carried out. On the other hand, convolutional neural networks have incomparable advantages over other methods in abstract information representation of two-dimensional images.
computer science, information systems,telecommunications
What problem does this paper attempt to address?